changingminds.org

How we change what others think, feel, believe and do

| Menu | Quick | Books | Share | Search | Settings |

Clicker Training

 

Techniques Conditioning > Clicker Training

How it works | Initial pairing | Reward natural action | Small steps forward | Cue last | Normalize | See also

 

Clicker training is a method for training an animal, usually a dog. The basic principle is to associate a reward with the clicker noise, then use the clicker into a cue for action.

While you may not be able to train people with clickers, there are important principles in here that may be useful!

How it works

The basic principle of clicker training is to first pair the clicker with a reward and then use the clicker as a reward substitute when training other actions.

Clickers work particularly well because:

  • They make a noise that has high frequency components that can be easily heard at a distance.
  • Nothing else sounds like the clicker so it easily gains attention.
  • The click sound is short, so it can be quickly heard and understood.
  • The sound of the clicker is consistent (unlike the human voice) and hence quicker to condition.
  • The click acts as a marker, indicating completion of a desirable action.
  • The click can be added while other things are going on and still stands out.
  • The clicker is quick and easy to use, so you can connect it immediately with desired actions.

A key feature of clicker training is that the click means one thing only, that a reward is coming. This means whenever you click, you must reward. The click says 'you did a good thing' at the moment of the good thing happening. The subject is hence more quickly conditioned as the action and reward are more clearly and closely connected in time. Note also that this means the click should never be used simply to get attention. It only means 'food is coming' and hence makes the dog feel good.

Clicker training often breaks a training session down into small steps, rewarding every small step along the way. The subject hence is incrementally guided rather than expected to complete the whole action before getting a reward.

This method requires strict discipline by the trainer. Clicks are only given after correct action. Reward always follows a click. Rewards are highly desirable. Punishment is never used.

A variation on clicker training is 'marker training', where some other signal is used instead, such as 'yes' or 'good dog'. The principles are otherwise the same.

The gentle approach of clicker training means it is suitable for use with puppies.

Initial pairing

The basic setup process is to associate the clicker with a reward ('priming' or 'charging' the clicker). If you are using food, then it is click-food, click-food. Get their attention first, click the clicker then give them the food so they quickly associate the sound with the delicious taste. This is typically done about 15-20 times in a session, anywhere from three to ten times a day. These are often done in short, five minute sessions.

Note that clicking means 'well done' not 'pay attention'. This means you need to get attention first by yourself. Ways to do this include:

  • Remove distractions such as toys or other people.
  • Move to a place where there are no distractions.
  • Select a time when the subject is not tired or otherwise unready to learn.
  • Do not wear clothes that may trigger reactions, such as a coat (=walkies!) or pyjamas (=time for bed).
  • Place yourself in front of them, occupying much of their field of view.
  • Get eye contact.
  • Show your pleasure at being with them.

When they do something desirable, push and release the clicker with a quick motion. It will usually click in both directions, with slightly different sounds. The more consistent you are with your push-release, the easier the subject will recognized and respond to it. Do not push it towards them and especially not their ears. Somewhere out of sight is good so it is only the sound that acts as the stimulus. If your subject runs away from the clicker, it may be painful for them and you may need some other sound (like making a clicking sound with your mouth).

You can also click and reward at random intervals through the day. This will help the subject to connect the click with pleasure. It will also teach them to pay attention to you just in case a click and reward might happen soon.

In order to maximise the speed and depth of association and consequent learning, the pleasure gained from the reward should be significant. For example, very tasty treats such as strong cheese or dried liver may be used for dogs. While making rewards very tasty, keep the size small. If you give a lot each time or something that they do not like that much, the attractiveness of the reward will wane and you will simply teach them that clicks are not worth bothering with.

Then click with desired action and finally the cue, as below.

Reward natural action

Start the training with something simple that the subject does a lot naturally. Natural actions include:

  • Sitting down.
  • Lying down.
  • Touching your hand with its nose.
  • Lifting a paw.
  • Moving in a desired direction (eg. coming towards you).
  • Following a target item you are holding (such as a spoon).
  • Being quiet when they might otherwise be noisy or agitated.

Encourage the desired action gently, without giving the command cue (which the subject probably will not understand). Hence it is better to wait until the subject is beginning to sit rather than pushing down its rear end. This can be helped by providing situations where it would naturally act as you want, for example when it wants to stay in one place for a while (such as when it realizes that paying attention to you is a good idea).

You can put food and clicker into your pocket and watch your subject across the day. When they do something desirable, such as sitting rather than jumping up or not barking at a visitor, then click and reward.

How you behave is also a reward or not. If you feel yourself getting frustrated, your body language will leak, so it is better to stop and cool down rather than get even a little irritated. Animals communicate with body language and can literally smell changes in mood.

Small steps forward

Use shaping to move gently towards the complete action, clicking and rewarding each incremental improvement. Be patient in this and be prepared for setbacks. As needed, step back to a reliable point and move forward again. Use gentle guiding if needed, though it is far better to catch the subject doing actions naturally.

An early step is to get attention, as described above. For this, click and reward when they respond to their name being called or even looking at you when you stand in front of them.

Ignore unwanted action. Paying attention to it can make it worse, distract the subject from desired action, or trigger them into fight-or-flight. Only click and reward desired action.

Luring can be used. For example holding a piece of food in front of the dog and moving it in the direction you want them to move (such as lying down). Click and reward initially when the start moving correctly, then later when they have completed the action.

To prevent something like biting or barking you may need some physical intervention. The moment the subject responds positively, stopping what they should not be doing, reward them.

Keep practice sessions short. Five minutes at a time is often enough. Watch for them getting bored and stop before they start losing attention. 

Cue last

Wait until the desired action is happening completely and reliably before adding the command cue (such as 'sit' or raised finger). The subject will then associate the cue with the complete action, not just the beginning of the move.

A good time to cue is when they start trying to get a reward by performing the action without any clicking. This is where they realize that completing the action is a good way to get reward.

Normalize

As the cue takes effect, you can fade out the clicker, gradually removing it. This may happen naturally as you give the cue and the subject automatically responds. It can help to replace the rewarding association of 'click' with something short like 'yes' or 'good dog'.

Make giving the command and getting the right response a normal, everyday activity that works well anywhere. Move from your normal training location to another similar place and eventually other places. Introduce distractions, from toys to other people. Introduce noise, such as from a busy street.

You can also teach them 'no', rewarding them when they stop unwanted action. Prevention training is a bit more difficult but uses the same principles, of reward only for correct action.

You may also want to train them to accept commands (or not) from other people, such as your family.

See also

Shaping, Fading

 

Site Menu

| Home | Top | Quick Links | Settings |

Main sections: | Disciplines | Techniques | Principles | Explanations | Theories |

Other sections: | Blog! | Quotes | Guest articles | Analysis | Books | Help |

More pages: | Contact | Caveat | About | Students | Webmasters | Awards | Guestbook | Feedback | Sitemap | Changes |

Settings: | Computer layout | Mobile layout | Small font | Medium font | Large font | Translate |

 

You can buy books here

More Kindle books:

And the big
paperback book


Look inside

 

Please help and share:

 

Quick links

Disciplines

* Argument
* Brand management
* Change Management
* Coaching
* Communication
* Counseling
* Game Design
* Human Resources
* Job-finding
* Leadership
* Marketing
* Politics
* Propaganda
* Rhetoric
* Negotiation
* Psychoanalysis
* Sales
* Sociology
* Storytelling
* Teaching
* Warfare
* Workplace design

Techniques

* Assertiveness
* Body language
* Change techniques
* Closing techniques
* Conversation
* Confidence tricks
* Conversion
* Creative techniques
* General techniques
* Happiness
* Hypnotism
* Interrogation
* Language
* Listening
* Negotiation tactics
* Objection handling
* Propaganda
* Problem-solving
* Public speaking
* Questioning
* Using repetition
* Resisting persuasion
* Self-development
* Sequential requests
* Storytelling
* Stress Management
* Tipping
* Using humor
* Willpower

Principles

* Principles

Explanations

* Behaviors
* Beliefs
* Brain stuff
* Conditioning
* Coping Mechanisms
* Critical Theory
* Culture
* Decisions
* Emotions
* Evolution
* Gender
* Games
* Groups
* Habit
* Identity
* Learning
* Meaning
* Memory
* Motivation
* Models
* Needs
* Personality
* Power
* Preferences
* Research
* Relationships
* SIFT Model
* Social Research
* Stress
* Trust
* Values

Theories

* Alphabetic list
* Theory types

And

About
Guest Articles
Blog!
Books
Changes
Contact
Guestbook
Quotes
Students
Webmasters

 

| Home | Top | Menu | Quick Links |

© Changing Works 2002-
Massive Content — Maximum Speed